1. Field of the Invention
The present invention relates to a replication system for implementing replication between a normal-system site and a standby-system site.
2. Description of the Related Art
Conventionally, computer systems have been used that are provided with a normal-system site (or an operating system site) and a standby-system site for maintaining functions even in the event of a disaster, such computer systems being referred to as “replication systems.” For example, the normal-system site normally operates to provide system functions, and the standby-system site operates in place of the normal-system site when the normal-system site is unable to operate normally.
In order to provide the functions of a computer system, the normal-system site and standby-system site each have storage unit for storing data. In a replication system, the data in the storage of the normal-system site are replicated and retained in the storage of the standby-system site such that the standby-system site is able to operate in place of the normal-system site, this process being referred to as “replication.”
In order to maintain the functions that were being implemented in the normal-system site when a difficulty occurs in the normal-system site in this type of replication system, the functions that were being carried out in the normal-system site are switched to the standby-system site.
A replication system has been proposed wherein, instead of transmitting data directly from the normal-system site to the standby-system site in the replication process, a relay transfer device is provided between the normal-system site and standby-system site and data are transmitted from the normal-system site to the standby-system site by way of a relay transfer device (refer, for example, to JP-2004-086721-A). The relay transfer device retains data either temporarily or permanently and transfers the data between the normal-system site and standby-system site.
Replication systems that use relay transfer devices include both synchronous systems in which the normal-system site and standby-system site are “synchronized” (hereinbelow referred to as “synchronous replication”) and asynchronous systems (hereinbelow referred to as “asynchronous replication”).
In synchronous replication, the writing of data to storage in the normal-system site prompts the writing of the same data to the storage of the standby-system site. In asynchronous replication, in contrast, writing to the storage of the standby-system site is not prompted by writing to the normal-system site but is carried out at a later time.
However, the above-described technology has the following problems:
In the event of a problem, it is essential that operations quickly switch from the normal-system site to the standby-system site, and moreover, that the functions of the normal-system site be reliably continued in the standby-system site to reduce any repercussions upon the functions of the computer system.
In the replication system of the prior art, however, synchronous or asynchronous replication from the normal-system site to the standby-system site takes time, and rapid switching was therefore problematic when switching after updating the data of the standby-system site. If switching occurs before completing data update in the standby-system site, data that preceded the update were accessed. Further, in order to guarantee that the most recent data are accessed, when accessing before completion of updating, an inquiry can be submitted to the relay transfer device to ascertain whether the object of the access is data that have been updated. However, such a procedure causes drops in performance such as an increase in the amount of processing and a reduction in throughput.
In addition, when a logical error occurs in the normal-system site, the replication process may cause the logical error to be reproduced in the standby-system site, preventing normal operation despite the switch to the standby-system site.
Finally, data of various levels of importance for the functioning of the computer system are included within retained data, and priority should be given to protecting data having greater importance when a problem occurs. However, data of high importance are in some cases not protected in replication systems of the prior art.